Debug your custom code in U-SQL/ADLA
By Xiaoyong Zhu
Created 1 year ago
Big data systems usually provide extensibility model through languages such as Java, C#, Python, etc., however the failures in the customized code are hard to debug, since some of the errors can only be reproduced during actual running time in the cloud. Due to lack of information during cloud run time, it is thus harder to debug run time errors for big data solutions in customized code. In the latest Azure Data Lake Tools for Visual Studio release, we have a feature called “Failed Vertex Debug”, where you can debug failed customized C# code using the same runtime and exact input data from the cloud. Developers can easily download the runtime and data which causes the code failure from cloud to local develop machine, run those resources locally to reproduce where goes wrong in the cloud, and be able to set breakpoints to see more detailed information. After that, developers can also iteratively debug the customized C# code, fix it, and then re-run the new piece of code in cloud.
Email this Mix